Energy Storage Device Control Method Based on Ensemble Empirical Mode Decomposition and LSTM

Abstract

The present invention relates to an energy storage device control method based on aggregated empirical mode decomposition and LSTM. Firstly, the historical short-term load data of the first n+1 time segments in consecutive n+2 time segments are normalized and aggregated. The subsequence or residual component obtained by empirical mode decomposition is used as the input item, and the subsequence or residual component corresponding to the historical short-term load data of the next period is used as the theoretical output item to train the LSTM model, and then the current time period and the distance from the current time The historical short-term load data of the most recent n time periods are preprocessed and input to the trained LSTM model, and then the trained LSTM model is used to output the predicted values, and all the predicted values ​​are reconstructed and denormalized The prediction result is obtained, and finally the charging and discharging of the energy storage device is controlled according to the prediction result. The method of the invention has high prediction precision, and the charging and discharging operation of the energy storage device is reasonable.

Description

technical field [0001] The invention belongs to the technical field of power load dispatching, and relates to an energy storage device control method based on ensemble empirical mode decomposition and LSTM. Background technique [0002] As an important part of the economic dispatching of the power system, accurate load forecasting can economically and reasonably arrange the start-up and shutdown of the generating units inside the power grid, ensure the stable operation of the power grid, and provide reliable information for power grid dispatching planning, equipment maintenance, and power grid reconstruction and expansion. data support. [0003] In recent years, with the continuous expansion of the field of electricity consumption, the number of users continues to increase, and the penetration rate of new energy in the power grid is getting higher and higher. Due to the intermittent and uncertain output of new energy, the peak and valley load of the power grid Therefore, on...

AI Technical Summary

Problems solved by technology

In view of the uncertainty of load forecasting, the above methods are often not ideal in terms of forecasting accuracy. With the rise of neural networks, their powerful learning ability and self-adaptive ability enable them to achieve great success in many fields such as pattern recognition, intelligent robots, and automatic control. Due to the uncertainty of the load, the neural network can be used for learning to improve the accuracy of load forecasting

[0004] At present, some people use wavelet analysis combined with neural network to predict time series, but wavelet analysis needs to select the appropriate mother wavelet and set the number of feasible decomposition layers. The adaptive effect of the decomposition of nonlinear and non-stationary signals is poor, and the prediction accuracy is still low. needs improvement

Images